►
Description
Many organizations find it difficult either to expand OpenShift or demonstrate business value from it. Today we will discuss ways to make integrating with other teams smoother. Whether this is as simple as having conversations to ask relevant questions about the developer and application teams’ requirements, or learning how to measure costs associated with OpenShift to prove value to the leadership, having conversations and supporting our peers in the dev and application roles is critically important to success as an administrator.
This week we’re joined by Christina Kyriakidou, author of the recent “As an OpenShift Administrator, what questions should you ask your developers” blog post, to talk about the technical and non-technical tasks and conversations administrators can take on to grow their presence in the company and show the value of OpenShift.
Want to learn more?
https://cloud.redhat.com/blog/as-an-openshift-administrator-what-questions-should-you-ask-your-developers?sc_cid=7013a000002qG1LAAU
Try Red Hat OpenShift: https://red.ht/3uVxkya
Follow us:
Andrew Sullivan https://twitter.com/practicalandrew
Jonathan Rickard https://twitter.com/jrocktx1
OpenShift https://twitter.com/openshift
Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_confirmation=1
About Ask an OpenShift Admin:
Do you have questions? We want to answer them! Every Wednesday at 11am ET, hosts Andrew Sullivan and Jonathan Rickard sit down to answer your questions about Red Hat OpenShift.
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3rYz5ZS
OpenShift on GitHub https://github.com/openshift/
#RedHat #OpenShift #Kubernetes #ContainerPlatform
A
C
B
B
Hello
and
welcome
to
this
week's
ask
an
open
shift.
Admin
office
hours
live
stream
here
on
red
hat
live
streaming,
my
name
is
andrew
sullivan
and
I
am
as
always,
joined
by
my
lovely
and
effervescent
co-host,
mr
johnny
rickard.
How
are
you
today,
sir?
I'm
good?
How
are
you
it's?
It's
a
beautiful
red
hat
day
johnny.
It
is.
B
It
always
is,
I
was
I,
you
know
I've
watched
that
intro
video-
I
don't
like
you
know,
50
times
or
something
how
many
times
have
we
been
doing
this
right.
So
how
many
times
have
we
had
that
video
and
I
was
just
noticing
how
much
thinner
I
look
in
the
very
ending
of
that.
I'm,
like
man,
has
their
last
year
really
been
that
rough
on
me,
jeez,
I'm
I'm
getting
older
over.
A
B
Real
so
anyways
yeah,
hello,
everyone
welcome
to
the
live
stream,
so
this
is
an
office
hours
live
stream
here
with
red
hat
live
streaming,
which
means
that
we
are
here
to
answer
any
questions
that
you
happen
to
have
anything.
That's
on
your
mind,
anything
that
you'd
like
to
talk
about
that
is
hopefully
openshift
related.
You
know
we
can't
help
with
with
other
things
we
might
try,
but
no
guarantees.
B
B
So
we
this
one
came
together.
It's
the
culmination
of
several
weeks
of
of
effort
of
coordination
with
our
guest
christina,
but
it's
a
topic
that
is
near
and
dear
to
our
hearts
and
something
that
is
is
interesting
and
exciting.
To
me
in
particular
somebody
who
has
been
held
many
roles
myself,
so
we
are
joined
today
by
christina
kiryakadu
nope.
I
messed
it
up.
Curia
aikido.
B
D
D
B
It's
that's
one
great
news
right,
you're
in
the
right
role
with
the
right
skills
right
and
doing
doing
something
that
you
enjoy.
But
two
is
something
that's
really
important.
I
think
a
lot
of
us
it
folks.
I
I
would
classify
classify
myself
as
an
introvert,
and
that
doesn't
mean
that
I
dislike
people
or
anything
like
that.
It.
It
just
means
that
sometimes
it's
it's
harder
to
get
that
started
and
christina.
B
You
can't
you
can't
just
throw
more
technology
at
the
problem
and
hope
that
it
sorts
itself
out.
You
know
we,
usually
it
was
in
the
days
of
cmdb
right
so
christina
you're
you're
in
the
uk,
you're
you're,
probably
familiar
with
the
oh
gosh.
What
is
johnny,
what
is
the
certification
we
all
had?
The
itil,
I
told
yeah.
C
B
All
one
word:
no
spaces
that'll,
redirect
you
to
a
page,
a
landing
page
that
has
the
recordings
and
the
slides
for
all
of
the
what's
new
and
what's
next
presentations
with
openshift,
so
you
can
go
ahead
and
see
all
of
those,
but
we
wanted
to
pull
out
a
couple
of
things
that
that
I
and
johnny
both
thought
were
particularly
interesting
and
christina.
You're,
of
course,
welcome
to
to
weigh
in
as
well.
B
If
there's
something
that
was
interesting
to
you,
so
the
the
first
one
that
I
want
to
poke
at
and
spend
just
a
couple
of
seconds
talk
about
talking
about
is
core
os,
layering
and
johnny
you
and
I
both
like
as
soon
as
they
said
this.
We
both
sent
messages
to
each
other,
saying:
that's,
really
cool
yeah,
so
there
was
a
blog
post
or
or
I
guess
it
wasn't
a
blog
post.
B
B
So
this
is
really
useful
for
all
those
use
cases
all
those
times
where
you
know.
Oh,
my
hardware
needs
this
special
driver
right.
I
need
to
add
a
driver
into
the
into
the
host,
or
my
security
team
is
insisting
that
we
have
this
package.
You
know
installed
to
the
host
and
you
know
all
of
those
kinds
of
things
you
know.
A
Yeah,
I
personally
think
that
this
is
a
great
feature
I
mean
like
just
it,
where
core
os
may
have
been
unapproachable
to
some
organizations
before,
because
you
couldn't
actually
modify
the
underlying
host
to
put
whatever
agent
you
had
to
have
on
your
host,
this
kind
of
solves
that
problem,
and
I
think
it's
it's
a
big
step
forward
and
I
think
it
shows
that
red
hat's
listening
to
some
of
the
pain
points
that
our
customers
have.
So
it's.
I
think
it's
pretty
awesome.
D
B
Yeah
it's
it's
funny.
You
say
that
and
that's
a
really
important
point
to
remember,
because
I
have
seen
a
lot
of
conversation
from
our
field.
Folks,
not
not
since
this
announcement,
but
previously
around
how
like
do.
We
really
have
to
support
rel
worker
nodes
like
that's
just
it
opens
up
so
much
possibility
for
people
to
just
accidentally
break
things.
B
B
Yeah-
and
I
should
also
also
remind
everybody
that
the
the
core
os
layering
and
the
next
couple
of
things
that
we'll
be
talking
about
are
they
are
roadmap
things
so
remember
that
they're,
not
shipping
today
and
technically,
there's
no
guarantee
that
they
will
ship
at
all.
It
is
roadmap.
So
just
keep
that
in
mind,
I
feel
like.
I
always
have
to
remind
folks
of
that.
A
Yeah,
I
think
that
the
big
one
for
me
was
the
just
the
overall,
like
the
the
whole
theme
of
everything
like
if
you
look
at
like
the
what's
next
from
last
time,
they
were
really
focused
on
like
just
security
across
the
entire
platform
right
and
then
this
time.
Now
it's
about
like
the
consistency
and
security.
So
now
there's
it's
a
lot
more
holistic.
So,
just
from
like
a
like
a
messaging
standpoint,
I
that
was
the
big
thing
that
stood
out
to
me
that
everything
that
they
talked
about
was
like.
Okay.
A
How
are
we
consistently
doing
this
across
the
platform?
How
are
we
like
integrating
security,
all
the
things
that,
like
us,
snaps
people
and
security
people
that
have
in
mind
it
was
kind
of
like
you
know,
checking
those
boxes
for
us
and
then
the
other
thing
I
thought
was
just
the
the
acs
stuff
was
pretty
awesome
like
where
acs
is
gonna,
be
a
cloud
service
or
it
could
be
potentially
cloud
service
and
then
the
dynamic
network
policies.
I
thought
was
that's
gonna.
A
D
C
D
B
Yeah
yeah,
interestingly,
there's
several
efforts
underway
to
create
some
our
solution
patterns
with
security
solutions
and
openshift,
so
johnny.
I
think
your
team
is
involved
in
one
of
those.
My
team
is
involved
in
one
of
those,
so
there's
there's
lots
of
stuff,
that's
coming
out
or
will
be
coming
out
in
the
not
too
distant
future
and
on
the
security
aspects,
so
keep
an
eye
on
all
the
various
red
hat
properties.
And,
of
course,
here
we'll
we'll
be
talking
about
it
as
well.
A
The
last
one
really
quick
before
we
move
on,
though
yeah
that's
gonna,
be
super
awesome.
Is
that
kcp
thing
I
you
brought
it
up
last
time
in
the
last
room.
That
is
just
the
more
I
look
into
that
the
more
rad
that's
just
gonna
be.
I
cannot
wait
for
that
to
happen.
That's
gonna,
be
it's.
I'm
super
excited
for
that.
Yeah.
B
B
The
underlying
you
know
open
shift
in
kubernetes
clusters,
away
from
the
user
from
the
developer,
the
application
team,
so
that
they
see
a
kubernetes
api
endpoint
that
they
have
full
control
over.
They
can
go
in
and
they
can
deploy
their
own
operators.
They
can
do
all
the
things
that
they
need
to
do
inside
of
there,
but
behind
that
api
could
be
one
or
ten
or
fifty
or
five
thousand
different
clusters,
and
it
just
transparently,
you
know,
works
for
their
workload.
B
B
Okay,
so
just
a
couple
of
additional
quick
things
that
I
wanted
to
hit
on
here.
Let
me
share
my
screen
as
soon
as
I
find
the
right
window,
nope
not
add,
source
we're
going
to
share
this
guy,
all
right,
a
couple
of
quick
things
so,
one
last
week
we
quickly
talked
about
how
so
there
was
a
upgrade
blocking
issue
with
fcd
in
4.8
to
4.9
upgrades,
so
that
has
been
resolved
with
4.9.28,
which
means
that
upgrades
in
the
stable
channel
are
now
available
again.
B
B
And
we
can
see
here
that
I
now
can
go
to
4.9.28.
So
for
anybody
who
was
waiting
for
that
issue
to
be
resolved,
you
can
now
update
your
clusters
as
a
part
of
the
stable
release
channel
so
yay.
I
guess
you
know,
plan
downtime
maintenance
windows,
whatever
that
happens
to
be,
and
then
we
can
continue
along
those
those
updates
that
should
also
or
that
that
does
not
mean
that
4.9
to
4.10
is
unstable.
B
Yet,
however,
so
if
we
come
down
here
to
stable
4.9,
then
I
want
to
go
to
my
target
version.
You
notice
that
there's
still
no
stable
for
4.10,
so
if
you
want
to,
if
you're
comfortable
using
fast,
you
can
do
those
4.9
to
4.10
updates
and
remember
that
4.8
to
4.10
is
the
first
update
that
will
have
the
the
accelerated
updates
associated
with
the
the
new
eus
policy
so
for
when
you're
doing
that
upgrades.
B
Basically
it
pauses
the
machine,
pull
machine,
config
pool
rollouts,
while
the
control
plane
does
the
4.84.9
4.10
upgrade
and
then
once
the
control
plane
is
at
4.10.
It
unpauses
the
mcps
and
the
nodes
go
directly
from
4.8
to
4.10.
So
and
we've
got
a
blog
post
about
that
somewhere.
I
need
to
dig
that
up.
B
B
B
B
So
if
we
scroll
down
here
they
have
they
link
into
the
trimaran.
I
think
that's
the
right
pronunciation,
I'm
terrible
at
all
languages,
so
my
apologies.
So
basically
this
defines
how
to
or
or
explains
how
to
add
the
secondary
scheduler
into
the
cluster
so
that
it
will
apply
to
any
of
your
workload
pods
right.
So
the
reason
why
it's
a
secondary
scheduler
is
because
the
regular,
the
the
primary
scheduler,
the
normal
openshift
scheduler,
still
applies
to
control,
plane
workloads,
but
the
the
secondary
one
applies
to
your
application
workloads
but
yeah.
B
Well
lots
of
cool
stuff-
that's
available
here.
If
you
actually
go
to
this
link
here,
you'll
notice
that
you'll
get
dropped
into
the
kubernetes
sigs
scheduler
plugins
and
from
here
there
is
a
whole
list
of
these
that
are
available
inside
of
here
right
capacity,
scheduling,
co-scheduling,
node
resource
topology,
so
try
moran,
there's,
there's
lots
of
these
inside
of
here.
So
if
it's
something
that's
interesting
to
you
check
it
out,
maybe
it
can
be
useful
for
your
particular
workload,
all
right.
So
with
that,
that's
that's
all
the
top
of
mind.
B
Topics
that
we'll
hit
on
today.
I
hope
not,
unless
you've
hit
the
bug,
fragmented,
etcd
updating,
is
not
going
to
happen
yeah.
I
understand
if
people
want
to
want
to
wait
a
little
bit.
You
know
pass
the
first
release
for
a
fixed,
rather
significant
bug
that
caused
a
midstream.
You
know
blockage
so
or
update
block.
I
get
that
so.
B
Yeah,
it's
funny
right,
because
we
we
always
talk
about,
and
we
always
remind
folks
that
the
fast
channel
is
ga.
It's
fully
supported.
There's
no
reason
why
you
can't
or
or
shouldn't
in
many
cases,
use
the
fast
channel
to
apply
updates,
especially
if
it's
something
that
is
important
or
relevant
to
you.
At
the
same
time,
I
was
an
administrator
for
a
long
time
right.
I
get
wanting
to
wait
for
stable
and
wanting
to
make
sure
that
it
has
been
thoroughly
tested
and
vetted.
B
You
know
in
order
to
make
sure
it's
not
going
to
accidentally
break
something.
You
know
the
the
engineering
folks
and
johnny
I
think
you've
you've
seen
christina,
maybe
as
well
right.
They
wrote
this
whole
big
long,
internal
blog
post
about
like
why
you
shouldn't
be
afraid
of
fast
and
why
you
should
be
using
fast,
and
I
I
I
think
internally
folks
recognize
that,
like
I
said,
I
recognize
the
value
of
fast,
but
I
also
very
much
empathize
with
with
you
all
with
our
audience.
D
D
C
B
B
Yes,
what
happens
when
you
deploy
a
new
core
os
node,
so
you
you
point
it
at
its
ignition
config
and
it
goes
to
the
machine
config
server
and
pulls
down
the
full,
the
full
set
of
machine
config
and
one
of
the
first
things
that
it
does
is
download
the
new
core
os
version,
which
is
in
a
container
and
put
it
onto
the
host.
So
technically
it'll
update
itself
to
the
current
version.
B
B
D
B
So
here's
the
change
log
and
there's
an
html
version
of
that
and
then
down
here
is
the
release.txt
and
very
much
to
your
point.
It
has
all
of
the
different
versions
everything
that's
inside
of
here,
including,
if
I
remember
the
name
of
it
remember
if
it's
height
yeah,
though,
and
to
be
clear,
there
won't
be
one
with
every
single
release,
but
there
should
be
a
container
image
inside
of
here
that
has
that
core
os
release
inside
of
it.
So
not
not.
Every
release
has
a
core
os
update,
yeah.
D
B
Yeah,
so
you
can
see
the
last
time
was
updated
was
the
fifth.
So
probably
since
we
do
weekly
releases,
it's
probably
two
releases
ago
and
in
case
anybody
didn't
know
so
the
way
to
re-lead
these,
these
releases
so
410.
So
this
is
the
openshift
version.
4.10,
8.4
or
84
is
the
rel
version
that
it's
based
off
of
so
8.4
in
this
instance,
and
then
this
is
the
date
that
it
was
actually
built,
so
20
204
05
at
0,
5,
41.,
yep.
B
D
B
B
D
And
you
can
use
it
with
in
combination
with,
like
advanced
cluster
management,
to
to
call
stuff
to
to
make
you
know,
automation
orchestration,
I
suppose
happen
in
fact
deployed
on
openshift
itself,
but
it's
because
of
the
vastness
of
of
the
topic.
I
don't
think
this
is
yeah.
This
is
the
right
forum.
B
C
B
There's
there's
lots
of
resources,
we'll
dig
those
up,
we'll
I'll
include
some
links
to
ansible
resources
in
the
blog
post
for
our
for
the
stream
all
right.
So
that's,
that's
all
the
all
the
top
of
mind
topics
we
have
for
today
for
the
audience.
Please
don't
hesitate
to
continue
to
ask
questions,
make
comments.
We
love
it
when
y'all
chat,
it's
it's
what
we're
here
for
so
please
don't
hesitate,
and
that
being
said,
let's
talk
about
today's
topic.
B
You
know
we
we
mentioned
gosh,
it's
been
what
six
weeks
ago
or
so
I
don't.
Let
me
see
here
where's
my
where's,
my
window.
I
think
I
brought
it
up
yeah,
so
here
christina,
your
blog
post
was
published
just
over
a
month
ago.
So-
and
I
think
we
mentioned
it
when
it
was
published
just
a
couple
of
days
after
that
and
it
triggered
a
conversation
between
us
of
like
hey.
This
is
really
great
information.
This
is
stuff.
That's
you
know.
B
B
It
would
be
really
great
if
you
could
join
us
on
the
stream
to
talk
about
not
only
what's
in
the
blog
post,
but
all
of
the
other
stuff
that
you've
got
going
on
because
you
are,
you
are
very
busy
right
and
and,
as
you
said,
you've
had
a
number
of
I
don't
know
about
roles,
but
certainly
a
number
of
different
areas
of
expertise
in
your
time
at
red
hat.
So
I
I
thought
it
would
be
a
phenomenal
opportunity
to
bring
you
on
and
talk
about
that.
D
Yeah,
so
so
writing
that
has
been
a
result
of
my
customer
asking
multiple
times
about
you
know:
how
do
we
do
this?
What's
the
best
way
of
doing
of
of
engaging
with
our
developers,
we
have
this
new
platform
going
on
and,
and
we
want
to
make
sure
that
we
ask
the
right
things.
Surely
surely,
red
hat
has
done
this
before?
D
What
do
you
have
when
it
comes
to
to
resources
to
make
that
happen,
and
it's
not
been
the
first
time
that
that
that
that
question
has
been
asked
and
it
kind
of
goes
to
show
that
if,
if
you're,
actually
being
really
lazy-
and
you
want
to
be
lazy,
you
should
write
the
blog
post.
The
first
time
that
this
has
happened.
B
D
Yeah
well
and
this
time
it
it's
I've.
I
think
it's,
the
third
fourth
time
that
that
this
has
been
asked
now.
At
that
point,
I've
had
enough
and
like
okay,
let's,
let's
write
it
down:
let's
collaborate,
let's
find
how
how
should
should
we
approach
the
subject
of
introducing
new
workloads
to
to
the
platform
and
my
my
customers,
the
ones
I've
asked,
have
got
it
right.
They
want
to
make
sure
that
you
know
they
increase
the
utilization
of
the
platform.
B
Yeah
it's
I've
said
before,
and
it's
worth
repeating.
I
think
that
you
pay
for
a
hundred
percent
of
the
platform.
Right,
especially
you
know,
in
a
hyperscaler
you're
paying
for
those
resources
right,
those
aws,
amis
or
azure
instances
or
whatever
it
happens,
to
be
you're
paying.
For
those
you
know,
open
shift
entitlements
and
all
that
and
if
it
sits
at
you,
know
two
or
five
or
ten
percent
utilized
that's
effectively.
You
know
resources
that
you're
not
using
effectively.
B
So,
of
course,
it
makes
sense
that
we
want
to
bring
that
utilization
up,
and
I
think
sometimes
as
administrators-
and
I
certainly
had
this
problem.
You
know
a
decade
ago,
15
years
ago,
when
I
was
a
storage
admin
when
I
was
a
virtualization
admin
and
we
were
just
kicking
off,
you
know
back
in
2007
2008
when
virtualization
was
just
starting
to
be
a
big
thing
like.
B
B
How
do
we
do
that
and
I
identify
right
help
to
understand
at
the
platform
level
what's
about
to
happen,
and
I
really
like
you
know
and
stephanie.
Thank
you
for
posting
the
link
into
the
chat
there.
So
I
really
like
some
of
the
the
questions
that
you
have
listed
here.
I
I
sometimes
I
was,
as
I
was
reading
your
blog
post
for
probably
the
fifth
or
sixth
time
just
before
this.
I
always
I
always
chuckle
at
the
core
platform
requirements
section,
because
I,
as
a
storage
admin.
B
B
Need
yeah
and
it's-
and
you
always
get
this
like
blank
deer
in
the
headlights,
look
from
the
application
developers
who
are
like
what's
an
iop
and
so
sometimes
with
resources.
It's
a
little
bit
of
a
guessing
game.
Unfortunately,
you
know
modern
platforms
are
flexible
enough,
that
we
can
make
mistakes
and
it's
okay,
but
I
I
less
concentrating
on
the
the
resource
requirements.
I
really
like
this.
B
So
I'll
pause
there
and
I'll
I'll
I'll
leave
you
with
a
question
christina,
which
is
you
know
from
your
perspective,
like
what?
What
is
that?
What
does
that
mean?
How
critical
is
that,
as
somebody
who
is
you
know,
practicing
this
day-to-day
interacting
with
customers
and
and
bringing
those
applications
on
board.
D
So
developer
onboarding
is,
is
paramount.
Actually
you
want
to
make
sure
that
you
have
the
process
from.
I
don't
know,
maybe
even
from
from
costing
to
getting
something
on
the
platform,
if
you're
an
enterprise
running
openshift
you're,
not
to
your
to
your
your
system,
administration,
that
does.
I
need
a
project
on
openshift.
It's
not
usually
how
it
works.
It
works
with.
I
have
a
cost
code
and
I
have
budget.
B
D
I
have
three
developers
that
want
to
work
online
on
on
on
containerization
on
this
particular
business
problem.
How
do
I
onboard
them
on
the
application
and
there's
quite
a
lot
of
things
that
you
you
want
to
do
and,
and
it
doesn't
necessarily
mean
that
you
have
all
the
answers,
but
what
you
should
start
with
is
automating
the
bits
that
you
have
the
have
know
the
integrations
for.
D
Cleanse
74.
it
likes
the
questions
in
the
blog,
but
loads
of
application
developers
can't
can't
answer
many
of
them.
Yes,
that
is,
that
is
entirely
correct.
They
cannot
answer
many
of
them,
but
it's
that
the
whole
point
of
having
those
questions
is
to
see
how
far
your
developers
are
during
their
design
phase
during
their
implementation
phase.
Have
they
considered
any
of
these
things
or
are
they
things
that
you
can
help
with?
So
if
they
don't
have
the
resources
for
consumption
for
these
things?
D
B
Yeah
when
we
were
chatting
yesterday,
you
know
about
the
stream.
We
talked
briefly
about
you
know
having
day
one
requirements
is
one
thing,
but
day
365
is
likely
to
be
entirely
different
than
day
one,
and
so
you
know
this
set
of
questions.
Is
you
know
it's
one
of
the
tools
as
you
go
through
and
do
that
onboarding
process
and
that
growth
process
and
that
learning
process
with
the
application,
the
application
team,
but
really
it
should
be
an
ongoing
conversation
as
both
sides.
Application
and
operations
learn
more.
D
Exactly
exactly,
and
you
say,
application
I
would
say
applications
so
that
the
the
first
thing
they're
going
to
start
with
is
one
use
case
and
one
set
of
requirements
and
to
say
I
would.
I
would
like
to
make
this
an
example
as
to
what
I
want
to
run
run
with
further
download
the
line
in
production
and
say:
I've
got
15
requirements
that
are
core
and
I
want
to
satisfy
them
and
then
on
on
day
180.
D
You
say:
okay,
well
what
about
application
number
two
and
then
you
say
by
the
end
of
the
year.
You
have
you
have
improved
your
applica,
your
platform
features
by
taking
in
account
what
your
application
teams
require.
So
if
I
say
as
application
one
I
have.
I
want
stateless
applications
running
on
openshift
and
I
want
a
little
bit
of
a
scale
up
and
down.
That's
that's
all
you're
gonna
get
there's
nothing
particularly
sophisticated
and
I
would
say
start
with
that
and
then
add
more
and
more
complex
features
on
it.
D
So
like
further
down
the
line,
you
might
want
to
say
I
would
like
to
add
storage,
replication
or
storage
clusters
attached
to
this,
so
that
I
can.
I
can
have
some
kind
of
disaster
recovery
across
the
board
for
applications
that
are
non-stateless
or
I
would
like
to
use
some
kind
of
a
backup
and
recovery
method
for
applications
that
require
it
again
that
are
non
that
are
stateful.
D
B
Yeah
yeah,
it's
it's
interesting
to
me
that
you
mentioned
some
of
those
things
and
I
say
that
there's
a
bunch
of
questions
we'll
get
to
those
in
just
a
moment.
Thank
you,
everybody
for
for
asking,
because
as
a
storage
admin
who
who
has
watched
and
participated
in
things
like
csi
and
seeing
how
storage
level
features
like
snapshots
like
replication,
you
know
we
we
had
annette
clewitt
on
a
few
weeks
ago
to
talk
about.
You
know
using
the
raman
dr
project
to
set
up
storage
replication
in
a
kubernetes
native
way.
B
B
You
know
I
look
at
you
know
when
I
was
a
storage
admin
we
had,
you
know,
features
like
deduplication
and
compression
and
snapshots,
and
all
these
other
things-
and
you
know
you
know
vmware
integration
with
you
know
the
storage
arrays
to
do
you
know
snap
offloaded,
snapchats
and
all
this
and
it
very
rarely
got
used,
and
I
think
it
was
largely
because
they
didn't
know
about
it
or
they
didn't
understand
it.
So
it's
interesting.
D
It
does,
and
it
doesn't
always
work
with.
I
don't
know
what
I've
seen
work
well,
is
having
discussion
forums
so
kind
of
having
two
hour
a
week,
perhaps
as
part
of
your
your
team
saying,
okay,
one
or
two
people
are
gonna
man,
this
and
and
we're
going
to
have
an
open
call
with
developers
to
solve
some
of
their
problems
for
the
people
that
are
running
on
the
platform
and
for
the
people
that
want
to
onboard
on
the
platform.
D
You
have
different
discussions
that
you
have
you
open
a
conversation
around
some
of
those
requirements
that
I
mentioned
in
that
in
that
blog
post
understanding
the
application
understanding
if
the
architecture
of
the
underlying
platform
would
satisfy
those
requirements
and
then
feed
those
in
to
your
product
owner
and
say,
I
would
like
to
add
these
feet
or
we
have
a.
We
have
demand
for
these
features.
D
Hopefully
the
product
owner
would
be
would
be
there
in
those
discussions
as
well
and
then
they'd
realize
we
have
a
demand
for
these
features
and
then
you'd
add
it
into
your
your
backlog
to
to
make
sure
that
they
happen.
But
as
you
develop
them,
you
would
want
to
make
sure
that
you're
doing
the
right
thing
for
your
developers,
because
you
know
a
new
feature
might
come
down
the
line
that
would
that
would
make
either
your
life
easier
or
much
more
difficult,
depending
on
on.
D
What's
going
on
so
so,
you'd
want
to
keep
either
in
you.
You'd
want
to
keep
the
developers
as
well
as
your
product
owner
in
the
loop
as
to
what's
going
on
in
your
platform,
development,
and
I
think,
having
interactivity
so
having
sandboxes
having
test
beds
for
your
developers
for
the
platform
is,
is
a
good
way
to
go.
B
B
I
was
just
looking
at
something
similar
today
of
how
to
extract
packages
like
so
you
can
get
the
the
precise
package
versions
that
are
in
a
specific
coreos
release
any
way
to
provision
a
pvc
with
local
storage,
and
thank
you,
john,
for
responding
host
path.
Provisioner
would
be
one
way
that
you
can
do
that.
B
So
there
is
the
gosh.
What's
it
called
now,
I
don't
think
it's.
I
think
it's
tech
preview,
or
maybe
even
like
just
upstream
available.
I
think
it's
the
the
local
local
storage
operator,
well,
there's
the
local
storage
operator
and
then
there's
another
one
that
is
related
to
what
the
odf
folks
are
doing
with
single
node
openshift,
where
it
it'll
provision
local
lvm
volumes.
C
B
A
number
of
different
ways:
one
thing
that
I
always
remind
folks,
is
that
with
the
host
path
provisioner,
it
doesn't
manage
the
underlying
paths.
So,
if
I
think
in
your
example
here
slash
a
slash,
app,
slash
app
if
that
doesn't
already
exist,
the
host
path-
provisioner
won't
create
it.
So
you
have
to
have
something
that
creates
that
partition
and
makes
it
available,
and
then
the
other
one
that
that
you
mentioned
there
johnny
the
local
path,
provisioner
that
one,
I
think,
there's
another
one.
B
B
B
B
B
Yeah
yeah
and
ort,
when
very
rightfully
right.
We
talked
about
that
when
he
was
on
the
stream
of
hey.
There
are
resource
requirements,
it
does
add
some
latency,
it
does
add
cpu
and,
and
you
know,
ram
and
network.
You
know
throughput
requirements
when
you
deploy
a
service
mesh,
so
take
that
into
account
lewis
newton
openshift-
I
don't
know
much
about
it.
Where
do
I
start
so
great
question,
one
that
we
that
we
love
getting
welcome
to
openshift?
We.
B
A
B
There
we
go
so
if
you
just
go
to
learn,
learn
dot,
open
shift,
cannot
type
and
talk
at
the
same
time.
So
learn.openshift.com
will
redirect
you
over
to
now
developers.redhead.com
and
if
you
scroll
down
here
there
is
a
whole
bunch
of
resources
that
are
available.
A
lot
of
them
are
developers
centric.
But
if
you
get
down
here
to
this
interactive
lesson,
section
there's
a
bunch
of
stuff
that's
inside
of
here.
B
So
a
lot
of
these
you
can
go
through
and
you
know
literally
like
the
openshift
4.9
playground,
you
can
click
the
start.
Button
and
it'll.
Take
you
to
the
screen
here.
I
click
start
and
oh,
there
must
have
been
one
already
ready.
Usually
it
it
waits
to
provision
one.
But
you
you
see
just
like
this,
I'm
in
a
real.
You
know
I'm
in
a
real
cluster.
B
B
B
B
I
realized
that
I'm
banging
this
drum
quite
a
bit
last
our
communities,
so
you
have,
among
other
things,
the
openshift
commons.
So
if
we
go
to
openshift.org,
which
now
redirects
over
to
okd.io,
so
our
upstream
commons
community
okd,
so
they
have
a
great
community
of
folks
to
help.
You
can
also
go
to
like
the
kubernetes
slack.
So
in
the
kubernetes
slack
there
is
an
openshift
users
channel
that
you
can
participate
in.
B
I
think
several
folks
who
regularly
watch
the
stream
are
active
in
there,
so
lots
of
different
ways
that
you
can
reach
out
and,
of
course,
here
on
the
channel.
You're
always
welcome
to
to
ask
questions
to
the
live
stream
hosts
our
hope.
Nine
discord.
Yeah
there
is
discord.
I
will
be
honest
and
point
out
that
I'm
not
very
active
on
discord,
because
too
many
chats.
A
There's
another
question
from
jamie:
I
mean
gonzalez
and
gonzalez
man.
I
I'm
sorry
if
I
ruined
that,
but
it's
it's
asking
if
we
can
speak
to
the
reasons
for
developing
on
the
platform
versus
developing
locally
and
then
running
on
the
platform,
so
why?
Why
should
I
develop
in
my
cluster
versus
local
and
then
push.
A
B
Yeah,
you
know
I
I'm
thinking
so
andrew's
getting
out
over
his
skis
here.
So
there's
concepts
in
the
developer
world
of
things
like
inner
loop
and
outer
loop,
so
inner
loop
is
being
able
to
quickly
write
and
test
and
iterate
on
your
code,
using
whatever
resources
so
a
lot
of
times
in
the
red
hat
portfolio.
We
talk
about
things
like
code,
ready,
workspaces
or
code,
ready
containers.
B
You
know
one
is
running
on
your
laptop
one
is
running
in
the
cluster
and
being
able
to
accelerate
that
inner
loop
activity
so
that
you
know
as
a
developer.
I
can
fix
the
bug.
I
can
add
the
feature
do
whatever
it
is,
and
then,
when
it's
done,
I
you
know
committed
to
get.
I
do
my
you
know
push
and
that
kicks
off.
B
You
know,
depending
on
a
number
of
factors
that
kicks
off
the
outer
loop
of
things
like
the
ci
cd
process
going
and
doing
the
builds
and
tests,
and
you
know,
code
reviews
and
all
that
other
stuff.
So
there
are
pros
and
cons
to
each
approach.
I
will
say
I
am
woefully
unqualified
to
talk
about
them.
Aside
from
I,
I
think
whether
it
happens
locally
or
on
the
platform.
B
D
And
also
when
you
develop
on
a
platform
when
you
develop
away
from
your
machine,
then
you
don't
have
the
it
works
in
my
environment
problem.
It
works
for
my
cluster
problem.
It
works
on
my
code,
ready
containers
infrastructure
because
you
may
have
permissions
you
may
have.
You
may
have
permissions
that
you
wouldn't
necessarily
have
on
the
actual
platform.
So
that
was
that's
one
of
the
things
that
I've
noticed
for
people
that
come
from
an
application
background.
D
I
want
to
do
stuff
on
a
more
secure
platform,
a
little
bit
more
restrictive
platform
that
things
don't
quite
work
the
same
way,
so,
I
would
say,
start
developing
in
an
environment
that
is
similarly
restrictive
to
production,
so
that
when
you
reach
a
point
where
you
are
in
production,
you
don't
have
to
change
many
things.
You
can
only
change
the
the
environment
of
variables
that
your
application
depends
on
rather
than
the
permissions.
The
underlying
infrastructure.
A
D
D
C
B
B
I
don't
ever
use
the
the
browser
client.
I
always
use
the
desktop
client,
so
man,
I
don't
know
how
to
extract
links
out
of
that.
So
I
think
it'll
work.
The
other
thing.
I
posted
a
link
to
a
lyft,
so
lyft
engineering
team
posted
a
series
of
blog
posts
about
how
they
scaled
their
application,
and
it
talks
a
lot
about
that
inner
loop
outer
loop
and
where
development
happens-
and
you
know
how
to
maximize
the
efficiency.
The
effectiveness
of
you
know,
developer
and
developer
time.
B
A
There's
one
more
question:
oh
I'm
sorry
yeah,
I
was
just
gonna
say:
there's
one
more
question
down
from
abba's
piece
about.
You
know:
they're,
building
their
private
cloud
with
openshift
in
the
lower
environment,
but
worrying
about
production
with
the
dr
setup,
so
basically
they're,
building
low,
deploying
high
and
so
they're.
The
concern
is
when
they
get
to
the
production
cluster.
How
do
they
set
up
dr
in
a
srm
type
fashion,
from
a
via
more
srm
type
fashion?.
D
D
B
B
We're
not
moving
the
vmware
environment,
we're
moving
the
vms
and,
if
you
think
about
it,
the
analogy
to
that
in
openshift
is
kind
of
what
you
just
talked
about
of.
Like
hey.
I've
got
my
application
definitions.
I've
got
my
data.
I
need
to
recreate
that
you
know
redeploy
my
applications
and
the
destination
you
know
having.
D
B
Well,
it
is
right,
and
certainly
with
something
like
openshift,
where
there
is
some
rigidity
in
the
expectations
that
it
has
from
the
environment.
So
the
analogy
I
used
the
other
day
is
you
know
again
I'll
pick
on
vmware
because
srm
you
know:
hey,
I'm
backing
up
the
operating
system
disk
in
every
one
of
my
esx
servers.
B
Okay
disaster
happens
at
the
primary
site.
Do
you
go
to
the
secondary
site
and
you
know
basically
restore
the
operating
system
disk
to
a
whole
slew
of
new
servers
turn
it
on
and
expect
everything
to
just
work.
The
same
way
it
did
before.
Well,
no
right
because
the
network's
gonna
be
different.
There's
different
dns
servers,
there's
different,
you
know,
maybe
maybe
they
aren't
exactly
the
same
physical
servers.
You
know
so
the
the
network
ports
are
in
a
different
order
right,
it's
it's.
D
So
I
would
say,
having
having
that
active
active
having
a
potentially
the
ability
to
to
deploy
to
an
environment
that
is
already
existing
kind
of
solves
some
of
those
problems,
because
you're
sure
that
your
dns
is
working
on
environment
too,
because
you've
pre-built
it
and
you've
tested
it.
Surely,
and
you
you're
sure
that
your
authentication
works
because
it
it
has
to.
D
C
B
It's
funny,
what's
like
the
number
one
troubleshooting
thing
like
when
somebody
says:
oh,
my
internet
isn't
working
and
I
just
had
my
son
literally
last
sunday,
my
son
dad
my
internet's
not
working.
I
don't
know,
what's
going
on
what'd,
you
do
to
the
internet,
why'd
you
break
the
internet
and
I
walked
upstairs
and,
and
I
can
see
the
back
of
his
computer
at
the
network.
Cable's
unplugged.
B
D
It
doesn't
even
have
to
be
a
full
you
can.
You
can
use
your
scalability
components,
openshift
has
and
says:
okay,
well
we're
going
to
keep
the
bare
minimum
here
and
once
we
have
a
bit
more
workflow,
he
can
it
can
scale
up.
But,
alternatively,
I
have
customers
that
are
running
applications.
Active
active
on
both
clusters
and
the
the
thing
that
that
keeps
them
together
is
a
is
a
gtm
at
the
top.
B
So
if
it's
a
worm,
dr
environment,
so
where
you
have
resources
deployed,
maybe
turned
off
right,
not
running
all
the
time,
but
importantly,
no
workload
deployed
to
it.
We
don't
require
entitlements
for
that,
so
you
can
turn
it
on.
You
can
apply,
updates
right,
keep
it
in
sync.
You
know
version
wise
and
stuff
like
that
with
the
primary,
but
until
you
deploy
workload
to
it
we
don't
require
entitlements
and
you
can
should
disaster
occur.
You
can
basically
take
the
entitlements
from
these
the
original
and
apply
them
to
the
destination.
B
So
any
questions
that
you
have
any
comments.
Please
go
ahead
and
submit
those
and
we'll
do
our
best
to
address
them,
but
so
one
of
the
things
that
we
had
talked
about
yesterday
and
as
we
were
leading
up
into
this-
is
how
do
we
as
administrators,
who
sometimes
you
know,
I
think,
struggle
to
understand
how
our
clusters
are
being
used
by
application
teams
and
stuff
like
that?
B
How
do
we
provide
that
business
value
justification
rationalization
up
to
our
own
management
chain,
and
this
is
you
know
it's
again,
something
that
I
when
I
was
a
storage
administrator
you
know
every
year
it
was.
Why
do
you
need
you
know
two
petabytes,
more
storage?
What
are
you
doing
with
the
other
30
that
you've
got
already?
You
know,
why
is
it
only
at
60
utilization
like
you've,
got
way
more
than
two
petabytes?
Why
can't
you
just
do
this
or
free?
B
D
Now,
what
we
have
is
the
cost
management
software
as
a
service
service
that
red
hat
provides-
and
I
find
I
find
that
it's
quite
useful-
to
use
that
to
report
what
you
have
currently,
of
course,
you
have
to
do
the
background
work
of
okay.
How
much
does
my
infrastructure
actually
cost
me?
What
is
the
management
cost?
What
is
the
server
cost?
What
is
the
actual
storage
cost,
etc?
To
run
my
underlying
infrastructure,
then
kind
of
once
you
have
that
set
down
you?
Can
you
can
say?
D
D
B
Yeah,
it's
it's
the
same
thing
in
my
mind.
I
I
map
it
to
the
same
thing
that
we
do
with
like
performance
right.
Oh
the
performance,
isn't
it
isn't
what
I
expect
it
to
be?
Here's
a
list
of
10
things
that
I
can
do
to
improve
performance,
and
you
want
to
kind
of
check
that
before
and
after
and
what
you
just
described
is
effectively,
you
know,
hey,
I
can
look
at
and
I
can
see
from
a
cost
perspective.
B
What's
the
most
efficient
way
of
running
my
clusters,
of
hosting
the
applications
in
my
clusters,
and
I
think,
there's
two
things
that
are
important
there,
one,
the
actual
cost,
doesn't
actually
doesn't
often
matter
right.
I
mean
yes,
I'm
sure
some
somebody
somewhere
cares
about
how
much
it
costs,
but
rather
it's
you
know
it
costs.
You
know
a
hundred
dollars
to
do
it
this
way,
but
I
can
prove
that
it
costs
ninety
dollars
to
do
it
this
other
way
or
a
hundred
and
ten
dollars
to
do
it.
B
This
other
way
and
it's
the
relative
cost,
which
brings
me
to
my
other
and
I'll
say
that
this
is
a
controversial
one.
I
don't
think
that,
like
you
were
talking
about
at
the
very
beginning,
identify
how
much
it
costs
in
infrastructure
and
in
manpower,
I
don't
think
you
actually
need
to
know
the
real
numbers
there
so
long
as
they
are
correct
in
their
relationship
to
each
other.
D
Well,
it
depends
entirely
on
whether
you're
doing
cross-charging
in
your
underlying
platform
but
you're,
absolutely
right,
because,
like
a
large
percentage
of
companies
using
openshift
kubernetes-like
environments,
don't
don't
necessarily
monitor
their
underlying
usage,
their
cost
and
you
can
learn
quite
a
lot
of
it.
So
from
it.
You
can
learn
that
I
don't
know
your
developers
prefer.
D
If
you're
doing
your
tagging
right,
you
developers
prefer
using
jboss
applications
or
they
prefer
using
python
applications
for
whatever
reason
over
something
that
your
organization
is
trying
to
to
to
promote.
Or
you
can.
You
can
learn
that
if,
if
you
do
you're
taking
across
cloud
providers
that
you
your
application
teams
like
using
a
one
cloud
provider
over
the
other,
when
you
have
a
clear
directive
to
use
something
else,
so
you
can.
D
You
can
learn
quite
a
lot
by
by
tagging
and
attacking
your
applications
and
using
cost
management
to
understand
a
little
bit
more
about
okay.
Where
does
your
cost
lie?
How
do
you
distribute
cost
across
cloud
providers?
And
if
you
do
your,
if
you
do
your
homework
right,
you
can
do
your
cross
charging
effectively
or
you
can.
Even
if
you
don't
care
about
the
prices,
you
can
learn
quite
a
lot
about
your
utilization
across
the
long
term.
A
Well,
it's
all
about
getting
value
right.
At
the
end
of
the
day,
it's
it's
getting
value
back
to
the
customer,
and
that
value
could
be
in
any
form.
It
could
be
financial,
it
could
be
performance,
it
could
be
availability,
there's
all
these
different
things
that
could
like
lead
into
like
that.
That
value
you
know
factor
or
whatever.
I
think
that
delta
solo
made
a
really
good
point
too
about
where,
like
a
lot
of
people,
think
that
on-prem
is
free
and
that's
just
not
the
case
right.
A
There's
like
there's
just
so
many
variable
things
that
go
into
power.
Space
and
cooling
data
center
costs
all
that
all
these
things
that
go
into
it
that
make
it
like
completely
the
opposite
of
free,
and
so
it
really
makes
a
good
argument,
though,
towards
cod
versus
on-prem
versus,
like
you
said,
azure
versus
aws,
like
I
prefer
this
over
this
because
of
the
you
know,
it's
just
the
little
nuance
things
that
devs
have
and
engineers
have.
C
B
Yeah,
how
many
of
us
have
those
spreadsheets
of
how
much
each
are
you
in
the
data
center
costs
and
and
how
much
admin
you
know?
How
much
does
an
administrator's
time
cost
and
how
do
I
aggregate
all
of
these
things
into
you
know
openshift
entitlements,
and
how
do
I
yeah
it's
it's
one
of
those
things
of
it's
never
going
to
be
right,
and
you
end
up
in
this.
You
know
pedantic
conversation
where
it
doesn't
matter
so
long
as
they're
relatively
close,
and
the
ratios
are
right
or
relatively
close.
That's
the
important
factor.
I'm
sorry.
D
So
I
was
going
to
say
at
that
point
where
you,
you
say:
bringing
value
to
customers.
I
know
organizations
that
keep
using
strict
quota
to
charge
their
customers
by
things
that
they
might
use
or
they're
up
to
the
point
that
they're
allowed
to
use
versus
to
the
things
that
they're
actually
using
and
that's
that's
the
point
where
you
actually
bring
value
to
your
customers.
Saying:
okay,
you've
used
the
entire
cluster
and
we're
very
happy
but
you're
paying
us
and
you've
used
only
a
tiny
bit
of
this
cluster
so
yeah.
D
This
is
the
the
respective
cost
of
of
what
we're
doing
or
you're
our
biggest
customer
and
therefore
we're
going
to
cater
for
your
features
now
because
you're
using
this
much
and
we
want
to
make
sure
that
you're
successful.
It
has
quite
a
lot
of
other
implications,
managing
and
understanding
where
your
costs
and
what
your
utilization
is.
B
Yeah,
I
I
I
think,
that's
something
that
I
don't
know
that
I
ever
learned
this.
While
I
was
an
active
part,
you
know
practicing
administrator
but
understanding.
You
know
what
your
customer,
you
know,
customer
being
the
application
teams,
the
developer
teams,
what
they're
actually
doing
with
the
platform
is
super
important.
I
was.
B
C
A
That
that
was
a
big
thing
when,
when
I
was
in
consulting
it's
like,
we
would
go
out,
and
you
know
we
deployed
these
clusters
and
like
there's
this
big
idea
like
we're
going
to
get
the
kubernetes
we're
going
to
containerize
our
application
and
then
what
happens
is
like
the
the
barrier
to
entry
was
a
little
high
and
you
know
if
they
didn't
have
somebody
kind
of
hand
holding
somebody
like
christina,
really
there
to
help
like
guide
and
like
all
right.
Look.
Here's
here's!
A
How
you
have
to
do
this,
here's
what
you
need
to
be
thinking
about,
then
they
might
get
the
hello
world
application
on
and
show
hey.
We
did
it
right
and
then
it
would
just
go.
It
would
go
flat.
My
friend
ko
he
would
sit
here.
He
talks
about
empty
clusters
all
the
time
like
that's
that's
like
his
main
mantra
and
getting
capability
on
these
clusters,
and
that's
really
the
thing
right
is
we
don't
want
these?
A
Yeah,
why
do
you
think
you
need
this
feature
that
that's
what
I
always
like
to
ask?
What
like?
What
do
you
think
you're
going
to
get
out
of
this
like
what?
What
exactly
is
is
the
outcome
that
you're
expecting
and
a
lot
of
times
what
they
think
is
going
to
happen
versus
the
reality
are
two
totally
different
things
and
so
yeah.
It's
definitely
important.
B
I
I
think
it's
it's
a
valid
example
right
service
mesh
is
it's
it's
complex,
it
does
add
value
and
our
hope,
nine.
I
like
your
comment
there
of
understanding
the
cost
of
something
versus
the
value
of
something
it
does
add
value,
but
you
want
to
add
it
because
you're
going
to
maximize
utilize
capitalize
on
that
value,
not
just
adding
in
something
for
the
sake
of
having
it
there
right.
If
you
build
it,
they
will
come
only
works
in
in
the
movies.
B
B
Yeah,
I
I
if
you
can
expand
on
that
a
little
bit
and
give
us
a
little
more
information
that
would
be
helpful.
You
know
openshift
is
used
by
telcos.
We
have
several
many
telco
customers
that
are
utilizing
openshift,
pretty
heavily
both
for
containerized
network.
You
know
cnv
containerized,
no
c
containerized
network
functions
cnfs
there
we
go
I'd,
get
there
eventually,
as
well
as
vnf's
through
tools
like
openshift
virtualization.
So
again
it's
the
use
case
thing
abbas.
B
If
openshift
can
auto
scale
the
pods,
when
marine
cpu
is
reached
threshold
helmet-
yes,
it
sure
can
so
that's
the
horizontal
pod,
auto
scaler,
as
well
as
the
vertical
pod
auto
scaler
can
both
do
that,
depending
on
how
you
want
to
scale
those
pods.
So
horizontal
basically
says
you
know:
hey
cpu
utilization
is
at
eighty
percent.
I'm
going
to
add
three
more
instances
of
the
pod
vertical,
auto
scaler
would
say:
cpu
utilization
is
at
eighty
percent.
I'm
going
to
now
add
an
extra.
B
B
Yeah
yeah,
thank
you
and
johnny.
You
mentioned
at
the
beginning
in
the
top
of
mind
keita,
so
keto
will
allow
you
to
do
that
based
off
of
other
arbitrary
metrics,
so
not
just
memory
and
cpu,
but
it
can
tie
into
things
like
prometheus,
and
you
know.
Oh,
the
web
server,
the
database
whatever
is
serving
x,
number
of
clients
or
x
number
of
records
per
second,
I
need
to
scale
that
up
or
scale
that
out.
A
And
one
thing:
that's
really
cool
about
keta
that
I
did.
It
was
like
the
thing
that
caught
my
eye
about
it
is
that
it
can
scale
to
zero.
So
it's
kind
of
got
like
that.
Serverless
function,
aspect
of
it
where
it
can
go
and
do
its
thing
and
then
scale
back
down
to
zero.
I
think
that's
gonna,
be
that's.
Gonna,
be
awesome.
C
D
Have
a
question
there
saying:
how
do
we
determine
the
split
between
helping
the
app
teams
restricting
their
ability
and
restricting
their
ability
to
do
their
work
as
platform
administrators?
You
will
know
if
you're
talking
to
your
development
teams,
you
will
know
that
something
is
not
working
for
them.
D
For
this
reason,
but
if
you
have
objections
around
this
particular
item
or
requirements
that
that
aren't
necessarily
meeting
but
being
met
because
of
it,
there
are
things
that
we
can.
We
can
work
around
so
and
you
should
be
able
to
do
that
if,
if
your
development
teams
and
your
business
applications
have
demonstrable
business
value.
B
B
We
used
to
hear
a
lot
about
shadow
I.t
and
how
it's
such
a
problem-
and
you
know
the
the
users,
the
you
know,
the
the
apps
teams
are
going
around
and
working
around
all
the
rules
that
we
have
in
place,
but
they're
using
a
corporate
credit
card
that
is
approved
all
the
way
up
the
management
chain
to
go
out
and
pay.
For
these.
B
You
know
shadow
I.t
resources,
which
means
that
it's
really
only
shadow
I.t,
because
you
know
only
consider
that
to
the
ops
team
or
the
infrastructure
team,
and
it's
only
because
they're
not
providing
the
service
that
the
business
needs.
So
I
I
you
know
today's
theme
right.
Having
that
that
open,
honest
conversation,
you
know
being
able
to
to
talk
about
you
know
what
are
the
problems?
How
can
we
bring
value
to
what
you're
doing
is
is
core
to
that.
D
I
was
gonna,
say
and
they're
saying
we're
getting
constantly.
We
can't
do
scenarios,
why
not
on
premise,
etc.
That's
a
conversation
you
can
have
with
them
and
then
say
to
your.
I
want
to
say
chain
of
command
management
chain
that
this
is
something
that
is
a
feature
that
is
a
requirement.
That
is
something
that
you
want
that
you
feel
is
necessary
because
you
have
these
customers
that
requiring
it.
D
B
If
we
do
this,
can
you
have
you
know,
can
you
can
you
make
the
application?
Do
this
right,
let's
meet
in
the
middle,
and
then
you
know
it
balances
out
that
that
complexity
that
cost
it's
it's
a
shared
responsibility.
Everybody
has
some,
you
know
some
vested
interest
in
it,
so
I
I
think
it's
something
that
gets
overlooked
a
lot.
I
don't
know
if
you
see
it
that
way,
christina.
B
That's
okay,
basically
just
saying
that
we
we
need
to
have
conversations
around
what
the
infrastructure
and
the
platform
is
capable
of
providing
versus
what
the
application
is
responsible
for,
whereas
previously
things
like
dr
the
application,
basically
absolved
themselves
from
it
right
srm.
I
just
I
put
it
in
my
vmware
s.
My
vmware
vm
and
srm
just
makes
sure
it's
at
the
new
site
and
we
can
make
openshift
behave
that
way,
but
it's
complex
and
it's
expensive.
So,
let's
meet
in
the
middle.
The
application
takes
on
some
of
that
responsibility.
D
B
Yep
that
whole,
you
know
devops
right,
just
open
the
communication
lines.
We
love
to
throw
around
the
term
devops
right.
That's
right,
all
right!
So
I'm
just
reading
through
any
of
the
comments
that
we
have
here.
I
do.
It
is
12
22,
my
time,
which
means
that
it
is
5
22
for
you
christina.
So
I
I
am
I'm
going
to
say
we'll
give
it
another
three
minutes
for
anybody
who
has
any
questions,
comments,
etc
to
weigh
in.
Please
don't
hesitate
to
do
that.
A
C
B
D
B
B
I
can
only
imagine
how
complex
it
is
to
actually
be
practicing
and
doing
underneath
something
like
that
delta
solo,
moving
away
from
dr
and
towards
application,
aha,
deploying
across
multiple
sites,
global
load,
balancing
yeah,
very
much
the
whole
cloud
native
application
thing.
It
doesn't
mean
that
our
we
don't
want
our
infrastructures
to
be
resilient
or
available.
B
Let's
see,
do
a
quick
overview
of
volume
snapshot
options
in
openshift,
I'm
not
quite
sure
what
you
mean
by
that
abbas,
so
volume
snapshots
generally
apply
to.
If
we're
talking
pvcs,
you
rely
on
your
csi
provisioner
to
do
those
snapshots
and
the
specific
implementation
will
be
up
to
the
you
know:
the
csi
driver
and
the
storage
vendor
that
you
happen
to
be
using
so
yeah
our
hope.
Nine.
Thank
you
so
yeah.
B
A
And
just
so
we're
making
stephanie
happy
and
to
key
in
on
delta
solo
about
you
know,
adding
this
as
a
reminder:
go
ahead
and
hit
like
the
the
like
or
subscribe
button
whatever.
It
is
in
youtube,
make
sure
that
you
tune
in
every
week
because
we're
all,
hopefully
we're
going
to
have
something
interesting
to
talk
about,
but
you
know
shameless
plug
for
our
show.
You
know.
Hopefully
you
come
out
and
watch.